**** BEGIN LOGGING AT Tue Apr 22 02:59:58 2014 Apr 22 05:00:43 khem: it's resolved now, but both wiki and patchwork were showing exception with mysql: "Exception Value: (1040, 'Too many connections')" Apr 22 05:47:43 JaMa: I think that box needs restart Apr 22 07:45:51 morning all Apr 22 07:46:10 hey bluelightning Apr 22 07:46:16 hi koen Apr 22 08:18:23 morning Apr 22 09:09:34 I remember there being a way to disable parallel make for do_install, but grep doesn't turn up anything Apr 22 09:09:50 for git I'd like to keep parallel make in do_compile, but run do_install with -j1 Apr 22 09:12:13 koen: PARALLEL_MAKEINST = "" Apr 22 09:13:26 hmm Apr 22 09:13:31 that var isn't used in meta/classes Apr 22 09:13:53 ah Apr 22 09:13:57 bitbake.conf magic Apr 22 09:14:13 yes :) Apr 22 09:15:06 thanks! Apr 22 09:23:32 i'd like to make bootable images with oe, the problem is that depending on the arch : bootloader, partitions and packages filesystems will differ, i don't think writing one recipe for every image is a good idea, and i don't think image.bbclass / image-*.bbclass / bootimg.bbclass has the flexibility needed to achieve this... am i wrong or i should write my own class to make an arch-agnostic image ? Apr 22 09:26:31 * kroon doesn't understand why the proposed fix for dbus-ptest-ptest in image, #5702, causes so much rebuilding Apr 22 12:03:23 Hi all. I am using openembedded but my question is graphics driver related. I was wondering if anyone could help me out. Apr 22 12:03:40 I am getting intermittently Error while setting position: X Error of failed request: Apr 22 12:04:05 and display doesn't wake up anymore until the restart Apr 22 12:15:12 just read about wic looks promising Apr 22 12:21:36 any1 Apr 22 12:21:36 ? Apr 22 13:50:46 afournier: the machine has variables which are obeyed by the images. machine and distro control what filesystems are emitted, and there are variables to add packages which are required to boot the hardware. partitioning is only supported in a limited fashion. machine, distro, and image are orthogonal axes, and must remain as such for optimum flexibility. generally speaking, any combination of machine, distro, and image can be used together, though Apr 22 13:50:46 there are exceptions Apr 22 13:50:47 * kergoth yawns Apr 22 14:05:26 kergoth: i agree, machine, distro, and image must remain orthogonal. then filesystem and partitioning should be orthogonal to that to :) it's just a bit confusing since oe gives the ability to choose the filesystem you want and at the same time relies on other tools like wic for more advanced filesystem management (like partitioning, mbr installation, etc.) Apr 22 14:06:37 partitioning and mbr installation would likely be doable through an image class, but non-trivial, just because of the limitations of our metadata format. trying to represent a given partitioning scheme via variables could be messy :) I can see why WIC is a nicer method for that Apr 22 14:07:47 agree. WIC looks like the way to go. Apr 22 14:31:41 is this bugtracker (https://bugzilla.yoctoproject.org) ok for oe-core related bugs and evolutions ? Apr 22 14:33:36 yes Apr 22 14:48:52 i've created a devicetree file for my board, what is the recommended way of getting it into the build? so OE generates a flattened file alongside the kernel image? Apr 22 14:49:31 send it to lkml and cc the devicetree list Apr 22 14:50:11 well, i would like to test it locally first :) Apr 22 14:51:16 It seems the kernel I'm using (linux-ti-staging) keeps its other DTS files in the kernel tree; does this mean I have to insert my own DTS file into the kernel tree as a patch? (basically, add the DTS patch to my kernel SRC_URI) Apr 22 14:52:11 yes Apr 22 17:24:18 http://www.linux.com/news/featured-blogs/200-libby-clark/770651-how-lg-took-webos-from-mobile-phones-to-tvs-in-under-a-year Apr 22 17:24:23 JaMa, +1 Apr 22 19:31:19 mago_: yes, patching it onto the tree is the way to go Apr 22 19:31:38 I suggest a bbappend in a personal layer :) **** ENDING LOGGING AT Wed Apr 23 02:59:58 2014